Three core areas of children's cognitive perceptions and behaviors that influence their mental health, academic achievements and wellbeing are 1) curiosity, 2) mindset and 3) their resilience to challenging environments. Curiosity is the cornerstone characteristic that promotes one's desire to learn and the intrinsic motivation to continue exploring in spite of obstacles that are ubiquitous in the learning process. A fixed or growth mindset has significant impact on children's academic and social achievements. Both have been shown to be malleable and susceptible to influence via social interaction. Finally, the embedding environment and the challenges it presents, serves to shape children's perception of their own capabilities as well as their skills. Social robot companions for children hold great promise in augmenting parents and teachers in the development, learning and promotion of academic achievements and wellbeing of children. While some of aforementioned concepts have begun to be explored in robotics modeling, a comprehensive expressive cognitive architecture that incorporates curiosity and mindset in a challenging environment, to achieve an autonomous, engaging and long-term robot companion for children does not exist. The long-term goal of the proposed project is to promote children's curiosity and growth mindset via interaction with a social robot companion. The project specific aim are to first develop and evaluate a novel, expressive, cognitive-affective architecture, by synergistically integrating models of artificial curiosity, understanding of mindsets, and expressive social behaviors to advance the state-of the-art of robot companions. The second specific aim is to evaluate the robot companion's influence on children's curiosity and mindset, by conducting a 6-month longitudinal study in which children will play and interact with the robot companion. The project aims lead to the development of a social robot companion that can both express and promote curiosity and growth mindset.